Search AI Products and News

After a nearly year-long pause, Meta has announced the return of its AI chatbot, Meta AI, to the European market. Starting this week, Meta AI will be rolled out across WhatsApp, Facebook, Instagram, and Messenger in 41 European countries and 21 other regions, but initially only with text-based chat functionality. Meta AI initially launched in the US in 2023. Although Meta had planned to bring the assistant to Europe, it was delayed due to concerns in Ireland.

Recently, the Meta AI team launched the video joint embedding prediction architecture (V-JEPA) model, an innovative initiative aimed at advancing machine intelligence. Humans can naturally process information from visual signals and recognize surrounding objects and motion patterns. An important goal of machine learning is to reveal the fundamental principles that drive unsupervised learning in humans. Researchers proposed a key hypothesis—the predictive feature principle—arguing that the representations of continuous sensory inputs should be able to predict each other. Early research methods utilized slow feature analysis.

As businesses increasingly adopt large language models (LLMs), improving the accuracy of model knowledge and reducing hallucination phenomena has become a significant challenge. Researchers at Meta AI have introduced a 'Scalable Memory Layer' in a recent paper, which may provide a solution to this issue. The core idea of the scalable memory layer is to add more parameters to the LLMs without increasing computational resources during inference, thereby enhancing their learning capabilities. This architecture is suitable for storing large amounts of factual knowledge while maintaining...

Large Language Models (LLMs) have made significant progress in the field of Natural Language Processing (NLP), shining in applications such as text generation, summarization, and question answering. However, the reliance of LLMs on token-level processing (predicting one word at a time) presents some challenges. This method contrasts with human communication, which typically operates at higher levels of abstraction, such as sentences or ideas. Token-level modeling also struggles in tasks that require understanding long contexts and may produce inconsistent outputs.

In recent years, large language models (LLMs) have made significant progress in the field of natural language processing (NLP), widely applicable in scenarios such as text generation, summarization, and question answering. However, these models rely on a token-level processing method that predicts word by word, which struggles with contextual understanding and often leads to inconsistent outputs. Moreover, when scaling LLMs to multilingual and multimodal applications, the computational costs and data requirements tend to be relatively high. To address these issues, Meta AI has proposed a novel approach.

Recently, Meta AI open-sourced a foundational multimodal language model named SPIRIT LM, which can freely mix text and speech, opening new possibilities for multimodal tasks involving audio and text. SPIRIT LM is based on a pre-trained text language model with 7 billion parameters, which has been continuously trained on text and speech units, expanding into the speech modality. It can understand and generate text like a large text model, while also being capable of understanding and generating speech, and even mixing text and speech to create various forms of expression.

Generating high-quality, temporally continuous videos requires a significant amount of computational resources, especially for longer time spans. The latest Diffusion Transformer model (DiTs) has made significant progress in video generation, but its reliance on larger models and more complex attention mechanisms has resulted in slower inference speeds, exacerbating the challenge. To address this issue, researchers at Meta AI have proposed a training-free method called AdaCache to accelerate video DiTs.

Recently, the Meta AI team introduced LongVU, a novel spatio-temporal adaptive compression mechanism aimed at enhancing the language understanding capabilities of long videos. Traditional multimodal large language models (MLLMs) face limitations in context length when processing long videos, and LongVU was created to address this challenge. LongVU operates primarily by filtering duplicate frames and employing inter-frame token compression techniques to efficiently utilize context length, allowing it to reduce video data while preserving visual details.

Recently, Meta announced on its WhatsApp channel that its AI chatbot Meta AI has officially launched in six new countries, including Brazil, the UK, the Philippines, Bolivia, Guatemala, and Paraguay. This marks a further internationalization of Meta AI, aiming to provide a smarter interaction experience for users worldwide. Meta also revealed that, in addition to these six countries, the company plans to gradually roll out Meta AI to more countries in the future, particularly in the Middle East. With this expansion,

Recently, a wave of 'Goodbye, Meta AI' has swept across social media platforms. Many users, including high-profile figures such as Tom Brady and musician Cat Power, have taken to Instagram to publish statements aimed at preventing Meta from using their data to train AI models. This phenomenon reflects users' deep concerns about data privacy and the application of AI technology, posing new challenges for tech giants in balancing technological innovation with user rights. Although these statements have no legal binding force, they highlight the growing sentiment among users.

At the recently held Meta Connect 2024 conference, Meta's founder and CEO Zuckerberg excitedly announced that Meta AI's monthly active users have nearly reached 500 million and is expected to become the most used AI assistant globally. During his speech, Zuckerberg mentioned, 'We are currently close to 500 million monthly active users and have yet to launch in some major markets, such as the European market.' This news undoubtedly fills people with anticipation for the future of Meta AI.

Meta AI has become the most used platform in India, receiving billions of queries just months after its launch, particularly excelling on WhatsApp. After overcoming cultural challenges and making algorithm adjustments, the AI now supports multiple languages and is available in 22 countries. The training computational requirements for the new model Llama4 are ten times that of its predecessor. Meanwhile, the user base for the Threads social app is nearing 200 million, and Facebook's usage among young adults in the U.S. is showing positive trends.

Meta's success with AI products in the Indian market is largely due to the widespread use of WhatsApp, which boasts 500 million users in India. Since their launch, AI products have processed billions of queries, demonstrating exceptional user retention and engagement. Despite challenges in adapting to local cultures, addressing sensitive religious topics, and mitigating image generation biases, Meta has adjusted algorithms to ensure fairness and supports multiple languages, including Hindi. Furthermore, Meta has announced it is developing the next generation of computing capabilities that are 10 times more powerful than Llama3.